price would depend on the investment regime. That is, for a given upstream fixed cost F¯, the wholesale price under the upstream investment regime would be larger than under the downstream investment regime 1wU pstream1F¯2> wDownstream1F¯22. Thus, a simple comparison of investment under both regimes would not be possible. Moreover, such an assumption would lead to an additional spot market distortion under the upstream in-vestment regime. While this would imply an interesting trade-offto analyze, it is beyond the scope of this study.
determining the investment regime, as different regulatory approaches may be optimal under different modes of competition. It has become a common notion in the literature to interpret different natures of competition as a manifestation of the importance of capacity constraints (compare, e.g., Tirole, 1988, and Kreps and Scheinkman, 1983). Moreover, the capital intensity of the upstream infrastructure also plays a crucial role as it influences the regulated wholesale price and hence, the monopolist’s margin, which is a major deter-minant of investment. As a central message of our findings, policy makers and regulators should be aware of the specific characteristics of the electricity sector under regulation when deciding on the optimal roll-out plan for smart metering devices.
Exclusive Retailing
3.1 Introduction
When new mobile handsets enter a market, they are often exclusively marketed through one of the mobile carriers in this market. Examples for such ‘exclusive retailing’ (ER) arrangements from the U.S. mobile phone industry include the introduction of LG’s
‘Chocolate’ via Verizon’s cellular service, Samsung’s ‘Instinct’ and Palm’s ‘Pre’ via Sprint, Blackberry’s ‘Pearl’ via T-Mobile, and Apple’s ‘iPhone’ via AT&T. Most of such exclu-sive retailing arrangements are abandoned once the handset receives recognition in the market. Market observers argue that ER is largely used by small mobile handset pro-ducers (see BCG, 2006)40. However, the rationale of the handset producers behind the adoption as well as the withdrawal of ER have not been formally analyzed. Their effects on competition and welfare have also not received due analysis.
In this chapter, we provide a rationale for such exclusive retailing arrangements. ER eliminates the disciplining effect of intrabrand competition between retailers, giving the exclusive retailer market power and hence, a higher retail margin. While creating such a double markup effect is costly for the manufacturer, it also comes with two profit-enhancing effects. First, it can serve as a mechanism to enhance brand-specific marketing investments by retailers (investment effect). This can be profitable for the manufacturer if the retailer owns a relatively more efficient marketing investment technology.41
Sec-40Examples outside the mobile phone sector include AC/DC’s new music album in 2008 which was exclusively sold via Walmart and the U.S. toy company Hasbro sells its toy action figures exclusively through Target (compare Subramanian, 2009).
41In our mobile handset context, this would mean that the ER arrangements enhance marketing invest-ments by the mobile carrier into the handset. BCG (2006) argues that ER is largely used by small mobile handset producers. These are often not that well known for their products and the mobile carriers have a much better distribution network as well as access to customer data. Hence, mobile carriers have often a relatively better technology in marketing mobile handsets. A similar argument can be made for a lot of other products, in particular when the manufacturer does not have direct contact to its final customers.
ond, ER might serve as a commitment device for reduced interbrand competition among manufacturers. That is, the elimination of intrabrand competition by one manufacturer leads to a unilateral price increase, incentivizing the competing manufacturer to increase prices (competition softening effect). While the investment effect can be interpreted as pro-competitive, the two other effects are clearlyanti-competitive.
Our model analyzes a vertically related industry where two upstream manufacturers pro-duce differentiated brands and sell them to two downstream retailers at linear wholesale prices. The downstream retailers resell the upstream goods to consumers. Producing as well as reselling the brand is associated with constant production/resale costs which are set to zero for simplicity. While consumers have different preferences regarding the brands, they are indifferent between retailers. The manufacturers can choose to sell their brands via both retailers or exclusively via one of the retailers.42 Thus, three different settings can arise: (i) Both manufacturers sell to both retailers. (ii) One manufacturer sells to both retailers, while the other manufacturer sells to only one retailer. (iii) Each manufacturer sells to only one retailer.43 Besides reselling brands, retailers can also con-duct brand specific marketing. Investment into marketing raises the perceived relative quality of a brand and enhances demand for the brand regardless of which retailer resells the good. We assume that the different parties cannot contract on a specific level of marketing effort.
When manufacturers sell their brand non-exclusively, intrabrand competition at the retail stage drives down prices to the retailers’ marginal cost, which consists of the wholesale price being paid for the brand. Retailers could invest into brand specific marketing. But as intrabrand competition eliminates any retail margin, retailers would never gain from such an investment and hence, do not invest. If a manufacturer adopts ER, intrabrand competition at the retail stage does not exist anymore. Hence, the retailer can gain a positive margin on the sale of the brand. Moreover, the brand specific investment becomes now retailer specific. Thus, investment becomes lucrative for the retailer and positive investment levels can be observed. In addition, the possibility of the retailer to gain a positive margin results in a double markup problem. This leads to an unilateral retail price increase of the exclusively sold brand and thus, weakens interbrand competition.
Using this model, we can derive the following results. First, we find that (wholesale and retail) prices, investment and retail profits are higher when one or both manufacturers adopt exclusivity relative to a situation where none of the manufacturers chooses exclusiv-ity. However, more exclusivity in the market does not necessarily mean higher equilibrium
For example, the microchip producer Intel partly relies on marketing by its downstream retailers.
42In other words, manufacturers can choose whether or not they want to have intrabrand competition at the retail stage.
43Here, we rule out the case where both manufacturers exclusively sell to the same retailer and hence, foreclose the second retailer from the market. Though it turns out that this setting would maximize industry profits, we believe that such foreclosure would never be allowed by competition authorities.
values.44
Second, with competition among upstream brands, we show that three types of equilibria exist, depending on the cost of investment and the intensity of interbrand competition:
(i) Both manufacturers do not adopt exclusive retailing (NER). (ii) One manufacturer does not adopt ER, while the other manufacturer adopts ER. (iii) Both manufacturers adopt ER. When adopting ER, manufacturers trade off the cost (double markup effect) and the benefits (investment effect and competition softening effect) of such behavior.
Equilibrium (i) occurs when the investment cost is large and interbrand competition is rather weak. Equilibrium (ii) occurs when both, investment cost and the intensity of interbrand competition, attain intermediate values. Equilibrium (iii) occurs for either very tough interbrand competition, very low investment costs or both.
Third, when upstream brands are asymmetric with respect to the investment cost in marketing their brand, the occurrence of equilibrium (i) is not affected. Moreover, the asymmetric equilibrium (ii) occurs for a larger parameter space, while equilibrium (iii) occurs for a smaller parameter space. In particular, for rather asymmetric brands the symmetric ER (iii) equilibrium only occurs for highly competitive markets.
Finally, we find that the incentive for a manufacturer to adopt ER contradicts with a welfare maximizing regulator’s view of ER. This gives scope for regulatory intervention.
In particular, ER should only be allowed if retail investment is sufficiently efficient or interbrand competition is rather tough.
This study contributes to two streams of literature, vertical restraints and exclusive con-tracting. Both analyze restrictions which are put on one trading party by another trading party. While in most of these articles the selling party restricts the buying party in es-tablishing alternative trading relationships, in this work the selling party commits itself to trade only with one of the buyers.
The literature on vertical restraints considers different kinds of restrictions such as ex-clusive dealing, resale price maintenance, exex-clusive territories, franchise fees and quantity forcing. Similar to our work, this literature models the vertical structure of an indus-try explicitly. Besanko and Perry (1993, 1994) model the equilibrium incentives to adopt exclusivity when interbrand competition exists. Though we also consider interbrand com-petition, they look at exclusive arrangements where a retailer is allowed to deal with only one manufacturer, while we look at situations where a manufacturer deals with only one retailer. While the former is most often described as exclusive dealing, we characterize the latter as exclusive retailing. Moreover, Besanko and Perry (1993) investigate the situation when a manufacturer can conduct retailer-specific investment. We, in contrast,
investi-44E.g., wholesale prices might be lower when both brands are distributed exclusively compared to a situation when only one brand is distributed exclusively.
gate the case when the retailer can conduct manufacturer-specific investment. Besanko and Perry (1994) do not consider investments at all, but foreclosure of retailers. We ex-plicitly rule out any foreclosure. Similar to our work, Mathewson and Winter (1984) and Winter (1993) analyze the role of vertical restraints for inducing retail advertising efforts.
They investigate the effect of different kinds of vertical restraints, but do not consider ER.
Moreover, they do not consider interbrand competition, as we do. Among others, Rey and Stiglitz (1995) discuss vertical restraints as a measure to weaken upstream competition.45 They consider a similar setting to ours, but restrict their analysis to symmetric outcomes only where either none or both manufacturers choose exclusivity, while we also analyze potential asymmetric outcomes. Armstrong (1999) and Harbord and Ottaviani (2001) analyze the link between the type of payment within a contract (lump-sum vs. linear payment) and the manufacturer’s decision to adopt exclusivity. They find that down-stream competition and the possibility of resale among retailers play an important role on the optimality of exclusivity. However, they neither consider investments nor upstream competition.
The literature onexclusive contracting considers only exclusive dealing arrangements and no other forms of restraints. In contrast to our work, this literature uses an incomplete contracting framework. Two different views of exclusivity arrangements can be found.
The anticompetitive view (Aghion and Bolton, 1987, Rasmusen et al., 1991, Bernheim and Whinston, 1998, and Segal and Whinston, 2000) argues that exclusivity serves to foreclose potential rivals from the market and hence, hinders competition. The pro-competitive view (Klein, 1988, Frasco, 1991, Marvel, 1982, Masten and Snyder, 1993, and Areeda and Kaplow, 1988) argues that exclusivity is needed in order to protect the return on non-contractible asset-specific investments. Without exclusivity investment incentives would be reduced. Such contracting arrangements can therefore be welfare enhancing, even when these lead to the foreclosure of potential rivals (compare, e.g., Fumagalli et al., 2009). Although we use a different methodology, we incorporate both the pro- as well as the anticompetitive view of exclusivity. Finally, de Fontenay et al. (2010) analyze the equilibrium incentives to adopt exclusivity in a modified Nash bargaining framework and apply the results to a prominent example of ER mentioned in our introduction, the
‘iPhone’ by Apple. Inc. However, they do not consider any anticompetitive elements of exclusivity.
In the next Section we present our model. In Section 3.3 we derive the outcomes for all different regimes of ER and in Section 3.4 the equilibrium incentives of manufacturers to adopt exclusivity. In Section 3.5 we analyze the incentives of manufacturers to adopt ER in case of asymmetric manufacturers. In Section 3.6 we derive welfare results and in Section 3.7 we conclude.
45Similar contributions, but with a different objective, are Telser (1960) and Jullien and Rey (2007).